Interpolation

Ryadia wrote:

There’s a page here on Interpolation if anyone is interested. Be careful it’s a 2 or 3 minute download with a 28k modem. Intended for broadband viewing is an understatement!

Ryadia

A URL is usually quite handy to go to a page…

—
Johan W. Elzenga johan<<at>>johanfoto.nl Editor / Photographer http://www.johanfoto.nl/

(Johan W. Elzenga) wrote:

A URL is usually quite handy to go to a page…

You were expected to manufacture that non-existent data by making your best guess, based on the information provided. This is Ryadia’s demonstration of how "interpolation" works.
——————————-
Tom

Unsolicited advertisements cheerfully ignored.

Tom Thomas wrote:

(Johan W. Elzenga) wrote:

A URL is usually quite handy to go to a page…

You were expected to manufacture that non-existent data by making your best guess, based on the information provided. This is Ryadia’s demonstration of how "interpolation" works.
——————————-
Tom

Unsolicited advertisements cheerfully ignored.

Geez Johan. What happened? You know so much about absolutely everything, I thought you’d just know where to go!

http://www.technoaussie.com/ryadia/

Have you tried "Genuine Fractals" ?
Fractal interpolation looks to me a superion method for interpolation, in theory IMHO.

http://hallogram.com/genfractals/

"Ryadia" a

Ryadia wrote:

Tom Thomas wrote:

(Johan W. Elzenga) wrote:

A URL is usually quite handy to go to a page…

You were expected to manufacture that non-existent data by making your best guess, based on the information provided. This is Ryadia’s demonstration of how "interpolation" works.
——————————-
Tom

Unsolicited advertisements cheerfully ignored.

Geez Johan. What happened? You know so much about absolutely everything, I thought you’d just know where to go!

http://www.technoaussie.com/ryadia/

While I agree with most of the technical aspect of what Ryadia’s article is saying, I think there’s just a bit of over-hyping going on here. Bicubic resampling followed by judicious use of unsharp mask has been standard operating procedure for decades. Advocates of this process has never claimed that they are creating actual image detail, but retaining the original detail without adding digital artifacting or "jaggies". It’s exactly like drawing on a balloon with a ballpoint pen, and inflating the balloon to make the image larger.

I remain skeptical of products such as "Genuine Fractals" because they do come close to claiming – without actually saying so – to be able to create image detail that is not otherwise available using standard resampling and sharpening. It’s as if the magic of fractals could clone self similar patterns into your image, and reveal depths of detail you never thought were there. To quote the Vancome Lady: "Know what? Uh-uh."

*BUT… at the same time, I agree with Ryadia that 300 or more pixels per inch is not necessary for an excellent image. An image printed at 100 pixels per inch can be a thing of beauty indeed, even for pixel sniffers. The trick is to have good contrast, color, and sharpening. Look how beautiful an image looks on a monitor at 72 or 93 pixels per inch. Why should paper, with its inferior dynamic range, require more data density than a CRT?

There is a megapixel frenzy now that was out of hand even at 3 megapixels. Look, it’s not uncommon for people to post questions to the camera groups asking whether their 4 megapixel camera will do acceptable 4×6 or 5×7 inch prints, or should they go to 8 megapixels! This is silly, and I think that some of the air needs to be let out of the marketing-hype-inflated megapixel balloon. To the extent Ryadia is promoting this view I agree completely.

So, "yes" – images may be printed much larger than most people say they can, and this idea needs to be promoted. And "no" – software such as Genuine Fractals does not create image data from whole cloth – merely interpolated numbers based on that data.
—

Mike Russell
www.curvemeister.com
www.geigy.2y.net

Your right.

in article ciqopj$19ib$, Pierre Chirouze at
wrote on 09/21/2004 7:42 PM:

Have you tried " http://www.artistmike.com " ? Looks to me a superion method for doing art, in
theory IMHO.

"Ryadia" a écrit dans le message de

Tom Thomas wrote:

(Johan W. Elzenga) wrote:

A URL is usually quite handy to go to a page…

You were expected to manufacture that non-existent data by making your best guess, based on the information provided. This is Ryadia’s demonstration of how "interpolation" works.
——————————-
Tom

Unsolicited advertisements cheerfully ignored.

Geez Johan. What happened? You know so much about absolutely everything, I thought you’d just know where to go!

http://www.technoaussie.com/ryadia/

I remain skeptical of products such as "Genuine Fractals" because they do come close to claiming – without actually saying so – to be able to create image detail that is not otherwise available using standard resampling and sharpening. It’s as if the magic of fractals could clone self similar patterns into your image, and reveal depths of detail you never thought

were

there. To quote the Vancome Lady: "Know what? Uh-uh."

I don’t have very good experience with the fractal method. Rezising with a factor of 2 was okay, factor 4 was looking very unnatural. A factor 2 can also be achieved with bicubic resampling 🙁

*BUT… at the same time, I agree with Ryadia that 300 or more pixels per inch is not necessary for an excellent image. An image printed at 100 pixels per inch can be a thing of beauty indeed, even for pixel sniffers. The trick is to have good contrast, color, and sharpening. Look how beautiful an image looks on a monitor at 72 or 93 pixels per inch. Why should paper, with its inferior dynamic range, require more data density than a CRT?

Not if you look closely (<30 cm) to the print. >250 PPI is definetely good enough. 600 PPI is too much.

A monitor is not comparable to a printer as a CRT does not need to do any halftoning. It can produce "all" colors for each pixel, where printers need to do screening with 4 or more colors.

But I agree on your story about the megapixels. You see a lot of cheap digicams with a huge amount of megapixels. Of which is not interpolated, the images are unsharp and contain a lot of noise.

Waldo

Pierre Chirouze wrote:

Have you tried "Genuine Fractals" ?
Fractal interpolation looks to me a superion method for interpolation, in theory IMHO.

http://hallogram.com/genfractals/

"Ryadia" a écrit dans le message de

Tom Thomas wrote:

(Johan W. Elzenga) wrote:

A URL is usually quite handy to go to a page…

You were expected to manufacture that non-existent data by making your best guess, based on the information provided. This is Ryadia’s demonstration of how "interpolation" works.
——————————-
Tom

Unsolicited advertisements cheerfully ignored.

Geez Johan. What happened? You know so much about absolutely everything, I thought you’d just know where to go!

http://www.technoaussie.com/ryadia/

The problem with Genuine fractals is not that it cannot do as well as other programs but that it takes a deal of messing around to actually enlarge an image. I make about 10 prints a day and if I used GF I’d be forever filling a hard drive with their native files which you have to have before you can actuall enlarge an image. Otherwise, I have had very good results with it on perfectly exposed images adn bloody awfull results with less than perfect pictures.

Ryadia

Mike Russell wrote:

Ryadia wrote:

Tom Thomas wrote:

(Johan W. Elzenga) wrote:

A URL is usually quite handy to go to a page…

You were expected to manufacture that non-existent data by making your best guess, based on the information provided. This is Ryadia’s demonstration of how "interpolation" works.
——————————-
Tom

Unsolicited advertisements cheerfully ignored.

Geez Johan. What happened? You know so much about absolutely everything, I thought you’d just know where to go!

http://www.technoaussie.com/ryadia/

While I agree with most of the technical aspect of what Ryadia’s article is saying, I think there’s just a bit of over-hyping going on here. Bicubic resampling followed by judicious use of unsharp mask has been standard operating procedure for decades.

*BUT… at the same time, I agree with Ryadia that 300 or more pixels per inch is not necessary for an excellent image. An image printed at 100 pixels per inch can be a thing of beauty indeed, even for pixel sniffers. The trick is to have good contrast, color, and sharpening. Look how beautiful an image looks on a monitor at 72 or 93 pixels per inch. Why should paper, with its inferior dynamic range, require more data density than a CRT?

There is a megapixel frenzy now that was out of hand even at 3 megapixels. Look, it’s not uncommon for people to post questions to the camera groups asking whether their 4 megapixel camera will do acceptable 4×6 or 5×7 inch prints, or should they go to 8 megapixels! This is silly, and I think that some of the air needs to be let out of the marketing-hype-inflated megapixel balloon. To the extent Ryadia is promoting this view I agree completely.
So, "yes" – images may be printed much larger than most people say they can, and this idea needs to be promoted. And "no" – software such as Genuine Fractals does not create image data from whole cloth – merely interpolated numbers based on that data.

I have used Bicubic resizing in 10% increments to enlarge some camera phone images with good results …(that is less than ideal results). Generally this process is worthwile when you don’t have any alternatives. The application I currently use allows me to upsize a file which I can then send to my printer without using Photoshop. For me this is a huge bonus.

Ryadia

I have used Bicubic resizing in 10% increments to enlarge some camera phone images with good results …(that is less than ideal results). Generally this process is worthwile when you don’t have any alternatives. The application I currently use allows me to upsize a file which I can then send to my printer without using Photoshop. For me this is a huge bonus.

Which application do you use? Above method is called "staircase interpolation" I believe.

Nice printer on your website. Are the prices of the consumables high?

Waldo

Waldo wrote:

I have used Bicubic resizing in 10% increments to enlarge some camera phone images with good results …(that is less than ideal results). Generally this process is worthwile when you don’t have any alternatives. The application I currently use allows me to upsize a file which I can then send to my printer without using Photoshop. For me this is a huge bonus.

Which application do you use? Above method is called "staircase interpolation" I believe.

Nice printer on your website. Are the prices of the consumables high?
Waldo

Interesting about the consumable prices Waldo. I use Canson paper in 100 metre rolls. It works out quite cheap. The ink is an odd thing. The reservoirs are all different sizes so they all run out together!

I used to get peeved at Canon and later Epson because I always seemed to be replacing ink tanks. Now I have a whole new bunch of factors. The ink is quite reasonably priced – compared to Epson’s pigment inks. However, under every silver lined cloud brews a storm.

The print heads are individual too, all 6 of them so while the ink might be priced well, every so often you have to put new heads in the colours. Cyan being the worst offender and this makes it no cheaper over a year than any other inkjet!

One thing I can say is that this printer is half the price of an Epson 4000, prints 6" wider than an Epson and has brighter colours whilst printing at the very least as good as my old 2200 Epson and many times much better. But it is slow.

Ryadia